Preparation and Properties of ${hbox {Mg}}({hbox {B}} _ {1-{rm x}}{hbox {C}} _ {rm x}) _ {2} $ Using Carbon Chemical Vapor Coated Boron

Abstract

Promising initial results on bulk Mg(B 1 x C x) 2 pre-pared with carbon doped boron are presented. Carbon doping is achieved by reaction of ethylene gas on boron powder using a stain-less steel tube furnace, a technique suitable for industrial scale pro-cessing. The nominal amount of doping was controlled by varying the reaction time with a fixed volume of ethylene gas, and the actual carbon uptake was determined by weight change after the reaction. The amount of carbon substitution x in the Mg(B 1 x C x) 2 was found using the angular shift in the (100) x-ray reflection. Carbon substitution by the full nominal content in the C doped precursor boron was obtained for doping up to 7.2at%, as shown by a -axis compression consistent with that of carbon doped single crystals. The critical current density of the 4at% C doped sample for tem-peratures at 20–30 K and fields up to 4 T, relevant to high tem-perature applications, was significantly higher than those in the published literature. The C of a 10wt% nano-SiC doped sample, used as a comparative benchmark, was found to be lower than the C doped sample at field below 2 T, but to reduce slower at higher fields. Structure analysis of the SiC doped sample revealed a coex-istence of two C substitution levels of 2.25%at and 5.25%at.

Full text

Page 1

2794IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 17, NO. 2, JUNE 2007
Preparation and Properties of Mg?B? ?C???Using
Carbon Chemical Vapor Coated Boron
E. A. Young and Y. Yang
Abstract—Promising initial results on bulk Mg?B?
pared with carbon doped boron are presented. Carbon doping is
achievedbyreactionofethylenegasonboronpowderusingastain-
lesssteeltubefurnace,atechniquesuitableforindustrialscalepro-
cessing. The nominal amount of doping was controlled by varying
thereactiontimewithafixedvolumeofethylenegas,andtheactual
carbonuptakewasdeterminedbyweightchangeafterthereaction.
The amount of carbon substitution x in the Mg?B?
found using the angular shift in the (100) x-ray reflection. Carbon
substitution by the full nominal content in the C doped precursor
boron was obtained for doping up to 7.2at%, as shown by a -axis
compression consistent with that of carbon doped single crystals.
The critical current density of the 4at% C doped sample for tem-
peratures at 20–30 K and fields up to 4 T, relevant to high tem-
perature applications, was significantly higher than those in the
published literature. The
used as a comparative benchmark, was found to be lower than the
C doped sample at field below 2 T, but to reduce slower at higher
fields. Structure analysis of the SiC doped sample revealed a coex-
istence of two C substitution levels of 2.25%at and 5.25%at.
?C???pre-
?C??? was
?of a 10wt% nano-SiC doped sample,
Index Terms—Boron powder, carbon doping, critical current,
MgB?superconductor.
I. INTRODUCTION
T
Currently NbTi is widely used in many superconducting appli-
cations, but requires complex and expensive Liquid He cooling.
A realistic MgB market will open-up if the critical current in
field,
(B), performance of MgB can be improved to a level
where it will compete with NbTi, with a lower cost and smaller
size of the cryogenics.
Doping with carbon has so far shown the most promising im-
provements in the field dependence of the critical current den-
sity
, for both wire and bulk MgB conductors, [1], [2].
The improvements in
are attributed to lattice defects in-
creasing the upper critical field
bution in the precursor with uniformity at nanometer scale is
therefore essential in order to take full advantage of bulk prop-
erty enhancements attributed to crystalline defects.
At present carbon has been added successfully by high
speed milling, [1] and grinding, [2]. However both methods
have the drawback of requiring precursor powder of a small
grain size, typically sub-micron nano-particles, which forms
HE 39 K transition temperature of MgB allows liquid-
cryogen free technology be used in the 20–30 K range.
. The method of C distri-
Manuscript received August 29, 2006.
The authors are with the Institute of Cryogenics, School of Engineering
Sciences, University of Southampton, Southampton SO17 1BJ, U.K. (e-mail:
e.a.young@soton.ac.uk; y.yang@soton.ac.uk).
Digital Object Identifier 10.1109/TASC.2007.897986
soft agglomerates due to the overriding effect of electrostatic
attraction at the nanometer length scale. Such agglomerates,
difficult to break mechanically, are the source of inhomogeneity
detrimental to achieving the uniform doping required. When
synthesized to MgB the dopants must be transported by diffu-
sion, hence require a higher reaction temperature that may not
be desirable for the optimal
the long milling times, required for good mixing, of the very
hard MgB grains with a micro-hardness around 1000 kgmm
comparable to that of SiC [3], significantly increase the risk of
contamination by the milling media.
In this paper, we describe a method of intrinsically homo-
geneous carbon loading by chemical vapor deposition (CVD)
of carbon on to the boron precursor powder. Such a method
presents two advantages. Firstly it allows the dopant to be dis-
tributed without contamination. Secondly the dopants are dis-
tributed evenly to every individual precursor grains so that a ho-
mogeneous nanometer scale distribution can be achieved with
a nano-boron powder. The realization of such a CVD carbon
loading using ethylene gas in a closed system is described in
the paper. Structure analysis was used to determine the level of
C substituted in bulk MgB samples prepared with the carbon
loaded boron powder by the liquid infiltration technique, [4].
Thecriticalcurrentdensityanditsfielddependence
alsomeasuredandcomparedwithdopingbySiCnano-particles.
The comparison firstly serves to highlight the improvements in
criticalcurrentrelativetoanotherdopant, for thesameliquid in-
filtration technique, and secondly demonstrates that doping by
CVD results in more homogenous crystalline properties.
performance. In addition
were
II. EXPERIMENTAL
A. Doping of MgB Precursor
A purpose built stainless steel tube furnace was used for the
chemical vapor deposition of carbon on boron powder. Ethy-
lene gas was introduced after the furnace tube was evacuated by
a turbo-molecular vacuum pump to 10
with a grain size of
, and a purity of 99.9% was treated
in ethylene gas for fixed periods. The mass of the boron powder
was measured before and after the process and the mass of
C deposited estimated from the mass gain. The powders were
then examined in a scanning electron microscope (SEM) for
morphology analysis. The precursor boron powder CVD coated
with carbon is referred as C-CVD boron in the following text,
and the corresponding samples as C-CVD MgB .
The nano-SiC doped boron was prepared by mixing the
powders together in methanol slurry under ultrasound for
several hours, then slowly heated up to 100
of methanol by evaporation.
mBar. Boron powder,
for the removal
1051-8223/$25.00 © 2007 IEEE

Page 2

YOUNG AND YANG: PREPARATION AND PROPERTIES OF2795
B. Bulk MgB Samples
With the C doped and SiC doped precursors MgB bulk was
made by the liquid infiltration technique, [4]. Boron pellets of
12mmdiameterand15mmthicknesswerepressedintoTalined
Fe crucibles. For each sample, a magnesium rod was placed on
the top of the pellet. Then the lid of the Fe crucible was closed
firmly above the Mg rod so that no air gap remained, and finally
thecruciblewassealedbyarcwelding.ExcessMgofabout10%
was used to compensate for any loss due to reaction with the Fe
and ensure there is sufficient Mg to complete the formation of
magnesium diboride. The Fe crucibles were placed in a vacuum
furnace and heated to 850
for 10 h then furnace cooled to
room temperature.
C. Characterization
The top face of the reacted pellets were ground back
to present a core surface for
(XRD) analysis. The magnetic hysteresis
cut with a diamond wheel to a typical dimensions of
, were measured in fields up to 9 T and temperatures
between 5 K and 30 K using a vibrating sample magnetometer
(VSM) in a Quantum Design Physical Property Measurement
System. The critical current density was determined using
, whereis the thickness of the slab.
3 mm
powder x-ray diffraction
of small slabs,
III. RESULTS AND DISCUSSION
A. Boron Powder CVD Coated With Carbon
On opening of the crucibles all samples were fully reacted
with no Mg rod remaining, and were removed as hard solid
pucks, without large cracks or other macro-scale defects.
Carbon deposition on the boron was confirmed by Energy
Dispersive Analysis, which showed no significant changes in
the powder morphology shown by Secondary Electron images
at a resolution of
.
B. Structure Analysis of C-CVD MgB
Due to a relatively low reaction temperature of 850
MgB pellets obtained were of high phase purity with some Mg
inclusions but no significant secondary MgB phases detected
by standard XRD analysis. X-ray diffractionof thereacted sam-
ples, (Fig. 1), shows increasing C doping shifts the (100) peak,
to higher angles, whilst the (002) peak position remains com-
parativelyunchanged.Bothpeaksbroadenconsiderablywithin-
creasing C content.
The change in the a-axis lattice parameter
the (100) peak, is plotted in Fig. 2 against the nominal doping
fraction x in the precursor power of Mg
latedfromthedepositedmassofC,alongsidedatafromtheliter-
ature[1],[2],[5].Fullcarbonsubstitutionatthenominaldoping
level is confirmed by the agreement of
MgB sample and C doped single crystal [5]. Therefore the
actual doping achieved in Mg B
inal doping fraction
in the precursor powder. The advantage
of the present method of C distribution by CVD coating of
B powder is evident when compared with mechanical carbon
doping by grinding [2] and high speed ball milling [1], which
onlyachievedpartialsubstitutionofthenominaldopingcontent.
, the
, derived from
B C , calcu-
between C-CVD
Cequates to the nom-
Fig. 1. Reflection of (100) from XRD of the C-CVD doped samples with in-
creasingnominaldopingpercentage ofC.Shiftofthe reflectiontohigher angles
correspondstoadecreasinga-latticeconstant(left),whilstthe(002)peak(right)
remains comparatively unaffected.
Fig. 2. The change in the a-axis lattice constant, ?? calculated from the (100)
reflection, against nominal doping fraction ?, from Mg?B
pression of ?-axis ?? of the C doped samples using C-CVD boron are in good
agreement with those of C-doped single crystals [5], indicating full substitution
of B by the nominal C content in the MgB lattice. Partial substitution obtained
with milling, [1] and grinding [2] are shown for comparison.
C ? . The com-
As no reduction in C substitution was found up to 7.2%at, the
technique looks to have the potential for even higher C doping
towards the theoretical saturation limit. The broadening of both
(100) and (002) reflections at higher C doping may be attributed
to strain and small grain size. It has been suggested that the lat-
tice disorder is a consequence of the C prohibiting the crystal
growth of the MgB [1].
C. Critical Current Density of C-CVD MgB
Fig. 3 shows the critical current density as a function of mag-
netic field at different temperatures between 5 K and 30 K. In
the log-log plot,
varies with
then reduces sharply with
fields. The transition field from one behavior to the other in-
creases with reducing temperature, e.g. 0.75 T at 30 K and 4.0
T at 5 K. Compared with other C doped MgB samples [1], [2],
at low fields and
or exponentially at higher

Page 3

2796IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 17, NO. 2, JUNE 2007
Fig. 3. Critical current density for 4%at C doped sample MgB
a function of magnetic field at temperatures between 5 K and 30 K incremented
steps of 5 K.
C
? as
Fig.4. Zero-field-cooledDCmagnetisationofsamplesmadefrompureBoron,
and Boron doped with 4% C-CVC, and 10wt% nano SiC doped.
the
at high temperatures and low/intermediate fields. In particular,
the
at 20–30K and in fieldsup to 4T is amongthe highest of
any MgB preparation in the published literature. A better per-
formance principally at high temperatures and lower field in-
dicates a clean C substitution without significant formation of
secondary phases, as also shown by the XRD data in Fig. 1. It is
also noted that the critical temperature
samplewasat36.6K(Fig.4),whichissignificantlyhigherthose
at 30–33 K for C doped single crystals samples [4], [5] with the
same a-lattice reduction of 0.02
surate nature of
with a-axis lattice parameter may be symp-
tomatic of a separation of structural shift and electronic doping,
and will be a topic of further investigation. At this early stage
in this work it is worth pointing out that the single crystals were
grownat1600–1900
,whilsttheC-CVDsamplesinthiswork
were reacted at 850
. It is possible that the reaction tempera-
ture affects nature of the C-B and C-Mg bonds.
of the C-CVD MgB samples exhibit is most improved
of the 4%at C-CVD
. The apparent non-commen-
Fig. 5. Comparison of critical current at 20 K for samples made from pure
boron, and boron doped with 4% C-CVD, and 10wt% nano SiC doped.
Fig. 6. The (100) reflection of the 10wt% nano-SiC, (symbols), deconvoluted,
reveals it is composed of two reflections, (dashed lines), the low angle is near
identical to the measured 2.25% C-CVD sample, (solid line), and the higher
angle peak can be interpreted from a linear interpolation of C-CVD doped data
in Fig. 2 as 5.25% C inclusion.
D. Comparison Between Carbon Doping by C-CVD Boron
and Nano-SiC
In Fig. 5 the
at 20 K of the 4% C-CVD doped MgB
sample is compared with that of 10wt% SiC doped sample,
with the un-doped control sample shown as the reference. The
C-CVD doped sample has a better performance throughout the
field range. It is noted, however, that the SiC doped sample
showed a slower
reduction between 0.5 T and 3 T, possibly
indicating contribution of a different pinning mechanism in this
field range.
It has beensuggested [1] thatimprovementsin
doping are due to C doping of the MgB . While the (100) re-
flection of the SiC doped sample shown in Fig. 6 is shifted to a
high angle as expected, the broad non-symmetric peak is signif-
icantly different from those found in the C-CVD samples. Two
underlying Gaussian peakswere found by
The peak at lower angle is very close to that of 2.3%at C-CVD
doping(soliddots),whilethehigheranglepeakcorrespondstoa
Cdoping percentageofabout5%atusingtheC-CVD andsingle
crystal data shown Fig. 2. The split of the (100) peak into two
peaks can be interpreted as an inhomogenous lattice structure
by SiC
fitting (solidline).

Page 4

YOUNG AND YANG: PREPARATION AND PROPERTIES OF 2797
with non-uniform C doping by SiC. Different C substitution
level in SiC doped samples is most likely the result of the in-
homogeneity in nano-SiC distribution in the precursor powder.
In contrast, the relatively clean XRD found in C-CVD samples
makes it ideal to identify the contribution directly resulted from
lattice substation by C, reducing the complications of a dirtier
system such as SiC doping. A broad two step superconducting
transition in SiC doped sample (Fig. 4) with an onset
as 36.4 K also underlines multiple C doping levels suggested by
the XRD data.
Further work to measure the
range of C-doping percentages will allow the possible interpre-
tation of the SiC doped sample into two parallel current paths
of different C doping levels.
as high
of samples made with a
IV. CONCLUSION
Initial results on the physical and magnetic properties of
MgB bulk made from C-CVD doped B. Full inclusion of C
has been demonstrated for samples of Mg B
0.04, and 0.07. At this stage the optimum percentage of C
for maximum
has not yet been found, but already the
C,
4% sample has the highest critical currents available in the
literature in the 20–30 K, 1–4 T range, where the first MgB
applications are currently anticipated. By comparing the XRD
and
properties of the C-CVD doped to a nano-SiC doped
sample, the dispersion of C is shown to be better by the CVD
route and the critical current consistently higher.
REFERENCES
[1] B. J. Senkowicz, J. E. Giencke, S. Patnaik, C. B. Eom, E. E. Hellstrom,
and D. C. Larbalestier, “Improved upper critical field in bulk-form
magnesium diboride by mechanical alloying with carbon,” Appl. Phys
Letts, vol. 86, pp. 202502–202504, 2005.
[2] W. K. Yeoh, J. H. Kim, J. Horvat, X. Xu, M. J. Qin, S. X. Dou, C.
H. Jiang, T. Nakane, H. Kumakura, and P. Munroe, “Control of nano
carbonsubstitutionforenhancingthecriticalcurrentdensityinMgB2,”
Supercond. Sci. Technol., vol. 19, pp. 596–599, 2006.
[3] S. M. Kazakov, J. Karpinski, J. Jun, P. Geiser, N. D. Zhigadlo, R.
Puzniak, and A. V. Mironov, “Single crystal growth and properties
of MgB2 and Mg?B? xCx??,” Physica C, vol. 408-410, pp. 123–124,
2004.
[4] G. Giunchi, G. Ripamonti, T. Cavallin, and E. Bassani, “The reactive
liquid Mg infiltration process to produce large superconducting bulk
MgB2 manufacts,” Cryogenics, vol. 46, pp. 237–242, 2006.
[5] S.Lee,T.Masui,A.Yamamoto,H.Uchiyama,andS.Tajima,“Carbon-
substitutedMgB2singlecrystals,”PhysicaC,vol.397,pp.7–13,2003.